Archives

Ramp Unveils AP Agents to Automate Accounts Payable

Ramp

Ramp, the leading financial operations platform, has unveiled Agents for AP, a suite of specialized AI agents designed to automate the most manual and time-consuming tasks in accounts payable, including invoice coding, approval, and payment processing. These agents integrate directly with Bill Pay, Ramp’s AP automation platform featuring industry-leading OCR technology, enabling invoices to be processed, approved, and paid in as few as three clicks.

“Ramp’s Bill Pay was already automating the majority of the AP process, and Agents for AP are now handling more complex workflows nearing 100% automation in certain cases,” said Karim Atiyeh, co-founder and CTO at Ramp. “This is only possible with agents that can source, understand and apply years of contextual knowledge to make decisions on crucial details like GL codes and approval recommendations.”

Agentic Workflow Automation for Accounts Payable

Agents for AP leverage the same context finance teams and approval stakeholders use to make decisions, autonomously connecting each invoice to vendor records, contracts, purchase orders, and approval history. Key capabilities include:

  • Invoice Coding – Agents apply logic from historical data and invoice details, such as product IDs and shipping addresses, to assign general ledger codes for each line item. With continuous learning, Agents now handle most invoice coding automatically, achieving 85% accuracy on accounting fields on the first pass and improving with each cycle.

  • Streamlined Approvals – Agents provide approval recommendations along with comprehensive summaries of vendor history, contracts, prior bills, and coding consistency, allowing approvers to make decisions quickly without additional research.

  • Card Payment Application – Agents identify card payment opportunities within vendor portals, eliminating manual data entry while capturing cashback benefits.

“For the finance team, Ramp has been a true game-changer. We can instantly analyze spend by vendor across all payment methods, and with the introduction of Agents and AI coding, we’ve eliminated tedious manual work while streamlining internal processes,” said Andrew Clarke, VP of Finance at STUDS.

Also Read: Fujitsu & Sony Bank team up to Embed Generative AI in Banking

Enhanced Fraud Protection

With 79% of businesses reporting payment fraud in the past year, Ramp’s AI agents provide critical safeguards against increasingly sophisticated attacks, including spoofed invoices, fake vendors, and unauthorized bank detail changes.

Early access customers using Agents for AP flagged over $1 million in fraudulent invoices in just 90 days. Agents analyze 63 data points, including payment history and vendor details, to detect suspicious invoices and vendor changes before payments are made. Ramp’s verification checks further confirm business identities, adding another layer of protection without extra effort.

Bill Pay Adoption and Growth

Since its 2021 launch, Bill Pay has become Ramp’s fastest-growing product outside of corporate cards, tripling payment volume and doubling customer adoption year-over-year.

“Ramp has completely modernized how we manage accounts payable. Instead of chasing down invoices, approvals, and vendor payments, everything flows seamlessly through one platform. Ramp has end-to-end automation, from invoice processing and approvals to payments and real-time ERP sync, which means we now close our books 2 days earlier and save hours of manual work every week,” said Muhammad Younes, Controller at Olipop.

Expanding AI Automation Across Ramp

Agents for AP follows the successful launch of Agents for Controllers, which automatically enforce company expense policies, eliminate unauthorized spending, and prevent fraud. Companies leveraging Agents for Controllers now automate 85% of expense reviews with 99% accuracy, detecting 15x more out-of-policy spend than traditional rule-based AI. Ramp plans to continue expanding its AI agent capabilities across its platform throughout the year.